GC-globulin (GC), or vitamin D–binding protein, is a multifunctional protein involved in the transport of circulating vitamin 25(OH)D and fatty acids, as well as actin scavenging. In the pancreatic islets, the gene encoding GC, GC/Gc, is highly localized to glucagon-secreting α-cells. Despite this, the role of GC in α-cell function is poorly understood. We previously showed that GC is essential for α-cell morphology, electrical activity, and glucagon secretion. We now show that loss of GC exacerbates α-cell failure during metabolic stress. High-fat diet–fed GC−/− mice have basal hyperglucagonemia, which is associated with decreased α-cell size, impaired glucagon secretion and Ca2+ fluxes, and changes in glucose-dependent F-actin remodelling. Impairments in glucagon secretion can be rescued using exogenous GC to replenish α-cell GC levels, increase glucagon granule area, and restore the F-actin cytoskeleton. Lastly, GC levels decrease in α-cells of donors with type 2 diabetes, which is associated with changes in α-cell mass, morphology, and glucagon expression. Together, these data demonstrate an important role for GC in α-cell adaptation to metabolic stress.
During metabolic stress, α-cell function becomes dysregulated, leading to inappropriate glucagon secretion and exacerbation of blood glucose levels (1), as well as impaired counter-regulatory responses (2). The mechanisms involved are multifactorial and include changes in α-cell glucose-sensing, α-cell dedifferentiation, paracrine feedback, hyperaminoacidemia, and α-cell mass (3–8). The importance of glucagon during metabolic stress and diabetes is exemplified by studies showing that deletion or blockade of the glucagon receptor protects against diabetes (9,10) and that immunoneutralization of glucagon restores glucose homeostasis in streptozocin-treated rats (11) or obese mice (12).
GC-globulin (GC), also known as vitamin D–binding protein, is an ∼58-kDa glycosylated α-protein that transports vitamin D metabolites and fatty acids in the circulation (13,14). GC is also among the most potent actin scavengers in the body and acts in concert with gelsolin to sequester actin filaments released from lysed cells (15,16). GC/Gc, the gene encoding GC, was thought to be almost exclusively expressed in the liver, where sterol derivatives, such as cholecalciferol, are converted into prehormone 25-OH vitamin D (25[OH]D) (17). However, recent studies have shown that human and mouse α-cells express equally high levels of GC/Gc (18,19), whereas the gene is absent from β-cells (19,20). Notably, GC constitutes an α-cell signature gene, since its promoter region is enriched for cell type-selective open chromatin regions (18). Moreover, a large-scale Mendelian randomization study of European and Chinese adults has revealed associations between GC single-nucleotide polymorphisms and type 2 diabetes (T2D) risk (21). Despite this, the tissue-specific mechanisms by which GC influences metabolic traits are poorly characterized.
Recently, we showed that GC contributes to the maintenance of α-cell function (22,23). Deletion of GC leads to smaller and hyperplastic α-cells that display abnormal Na+ conductance, Ca2+ fluxes, and glucagon secretion. This effect was found to be mediated by changes in the F-actin cytoskeleton, with a large increase in microfilament thickness and density in GC−/− islets. Notably, up until this point, the actin-binding properties of GC had only been described in the circulation. These studies were recently corroborated by Patch-sequencing (seq) studies, where GC was shown to inversely correlate with peak Na+ current in human α-cells (8). Despite the role of GC in maintaining α-cell identity, how GC impacts α-cell responses to metabolic stress remains unknown. Studies have suggested that manipulation of GC levels is able to restore glucose homeostasis in obese mouse models, although the exact cellular mechanisms are unclear as GC is upregulated in dedifferentiated β-cells (24). Although GC is an α-cell signature gene, it is also unclear whether abundant circulating levels of GC may also impact glucagon secretion and α-cell function.
In the current study, we set out to understand the role of GC in α-cell function during metabolic stress. By combining GC−/− mice and pancreas sections from T2D donors with immunohistochemistry, live imaging, and high-resolution microscopy, we show that GC is required for α-cell function and glucagon secretion under diabetogenic conditions. Pertinently, α-cell dysfunction can be rescued by restoring GC levels using exogenous protein.
Research Design and Methods
No data were excluded unless the cells failed to respond to positive control, responded inappropriately to negative control, or displayed impaired viability. All individual data points are reported. The measurement unit is the animal, batch of islets, or pancreas section, with experiments replicated independently. Animals and islets were randomly allocated to treatment groups to ensure that all states were represented in the different experiment arms. Animals and pancreas sections were coded to allow blinded analysis.
GC−/− mice harboring deletion of exon 5 of the GC gene were backcrossed to C57BL/6J for 10 generations (25). These mice have undetectable circulating GC as well as 25(OH)[3H]D3 binding (22,25). Mice were housed in a specific pathogen-free facility with ad lib access to food and water. Vitamin D sufficiency was ensured by using chow supplemented with 1,000 units/kg cholecalciferol. Mice were fed standard chow (SC) or a high-fat diet (HFD) containing 60% fat (Research Diets, cat. no. D12492), and body weight was checked weekly from 0 to 12 weeks. Male and female mice were fed the SC or the HFD from 8 weeks of age (numbers reported in the figure legends).
Formalin-fixed paraffin-embedded pancreas sections were obtained from the Alberta Diabetes Institute IsletCore. Quality control and phenotyping data are available for each preparation via www.isletcore.ca.
Glucose and Insulin Tolerance Tests
Mice were fasted for 4–5 h (8:00 a.m.–12:30 p.m.) before an intraperitoneal injection of 1.5 g/kg of sterile filtered d-glucose. Tail vein sampling was performed at 0, 15, 30, 60, 90, and 120 min postinjection. Glucose levels were measured using a Contour XT glucometer (Bayer). For plasma glucagon measures, mice were fasted for 4–5 h (8:00 a.m.–12:30 p.m.) before an intraperitoneal injection of 0.75 units/kg of insulin (Humulin S, Lilly). Blood was collected at 0 and 30 min postinjection and stored at −80°C pending ELISA for serum glucagon (Mercodia, cat. no. 10-1281-01). Values lower than the assay detection limit were interpolated from the standard.
Islet Isolation and Culture
Mice were humanely euthanized by rising CO2 and cervical dislocation (Schedule 1, Animals [Scientific Procedures] Act 1986, U.K.), before bile duct injection and inflation of the pancreas with 1 mg/mL SERVA NB8 collagenase (AMSBIO, cat. no. 17456.02). Islets were purified using Ficoll-Paque (Cytiva, cat. no. 17144003) or Histopaque (Sigma-Aldrich, cat. no. 11191, cat. no. 10831) gradient separation and maintained at 37°C and 5% CO2 in RPMI 1640 medium (Gibco, cat. no. 21875034) containing 10% FCS (Sigma-Aldrich, cat. no. F9665), 100 units/mL penicillin, and 100 μg/mL streptomycin (Gibco, cat. no. 15140122).
Immunostaining of Mouse Pancreases
Pancreases were harvested, incubated overnight with 10% formalin, dehydrated, and embedded in wax. Sections were cut at 5 µm using a Leica microtome before deparaffinization and blocking with PBS-Tween + 1% BSA for 1 h. Sections were incubated with primary antibodies overnight at 4°C before washing in PBS-Tween + 0.1% BSA (Sigma-Aldrich, cat. no. A3803) and incubation with secondary antibodies for 2 h at room temperature.
Primary antibodies used were rabbit anti-insulin, 1:500 (Cell Signaling Technology, cat. no. 3014, RRID:AB_2126503), mouse monoclonal anti-glucagon, 1:2,000 (Sigma-Aldrich, cat. no. G2654, RRID:AB_259852), mouse anti-somatostatin. 1:1,000 (Thermo Fisher Scientific, cat. no. 14-9751-80, RRID:AB_2572981), and rabbit anti-DBP, 1:1,000 (Sigma-Aldrich, cat. no. HPA019855, RRID:AB_1849545). Secondary antibodies were goat anti-rabbit Alexa Fluor 633 (Thermo Fisher Scientific, cat. no. A-21052, RRID:AB_2535719), goat anti-rabbit Alexa Fluor 488 (Thermo Fisher Scientific, cat. no. R37116, RRID:AB_2556544), goat anti-guinea pig Alexa Fluor 488 (Thermo Fisher Scientific, cat. no. A-11073, RRID:AB_2534117), goat anti-mouse Alexa Fluor 488 (Thermo Fisher Scientific, cat. no. A-11029, RRID:AB_138404), and goat anti-guinea pig Alexa Fluor 568 (Thermo Fisher Scientific, cat. no. A-11075, RRID:AB_2534119), applied at 1:1,000. Specificity of antibodies was based on known cell type colocalizations, overlap with insulin, glucagon, or somatostatin reporters, or loss of staining in knockout tissue. F-actin and G-actin were visualized using Phalloidin-iFluor 488 (Abcam, cat. no. ab176753) or DNaseI-594 (Invitrogen, cat. no. D12372), respectively.
Imaging was performed using Zeiss LSM780 or LSM880 meta-confocal microscopes, equipped with sensitive gallium arsenide phosphide (GaAsP) spectral detectors. Alexa Fluor 488, Alexa Fluor 568, and Alexa Fluor 633 were excited at λ = 488 nm, λ = 568, and λ = 633 nm, respectively. Emitted signals for Alexa Fluor 488, Alexa Fluor 568, and Alexa Fluor 633 were detected at λ = 498–559 nm, λ = 587–629 nm, and λ = 637–735 nm, respectively. Superresolution images of F-actin were acquired using a Nikon N-SIM S microscope, SR HP Apo TIRF 100× 1.49 numeric aperture (NA)/oil immersion objective, and ORCA-Flash 4.0 sCMOS camera, with online deconvolution. Alexa Fluor 488 and Alexa Fluor 568 were excited at λ = 488 nm and λ = 568 nm, respectively. Emitted signals were detected at λ = 500–550 nm and λ = 570–640 nm for Alexa Fluor 488 and Alexa Fluor 568, respectively.
Intracellular Ca2+ Imaging
The ratiometric Ca2+ dye, Fura2 (1 mg; HelloBio, cat. no. HB0780), was loaded into islets using 20% pluronic acid dissolved in DMSO (Thermo Fisher Scientific, cat. no. P3000MP) at 37°C for 40 min. For islets treated with GC, 5 μmol/L GC (East Coast Bio, cat. no. LA166) was added during Fura2 incubation. Islets were transferred to the heated chamber (34°C) of a Nikon Ti-E microscope coupled to a 10×/0.3 NA air objective (Nikon Plan Fluor), allowing simultaneous cell resolution imaging of multiple islets (lateral resolution = 910 nm). A Cairn Research FuraLED system provided excitation at λ = 340 nm and λ = 385 nm. Emitted signals were captured at λ = 470–550 nm using a Photometric Delta Evolve electron-multiplying charge-coupled device. Intracellular Ca2+ levels were shown as the emission ratio at 340 nm and 385 nm. Several experiments were repeated using the nonratiometric Ca2+ dye Fluo 8 (AAT Bioquest, cat. no. 20494). Confocal excitation was delivered at 470 nm (emission λ = 500–550 nm) by a North 89 LDI Illuminator, CrestOptics V2 X-light spinning disk and 20×/0.75 NA air objective (Nikon Plan Apo λ). Intracellular Ca2+ levels were quantified as F/Fmin, where F is fluorescence at any given time point, and Fmin is mean minimum fluorescence. All experiments were performed in HEPES-bicarbonate buffer containing (in mmol/L) 120 NaCl, 4.8 KCl, 24 NaHCO3, 0.5 Na2HPO4, 5 HEPES, 2.5 CaCl2, 1.2 MgCl2, and supplemented with 0.5–17 mmol/L d-glucose.
Insulin and Glucagon Secretion Assays
HEPES-bicarbonate buffer was used for all assays, containing (in mmol/L) 120 NaCl, 4.8 KCl, 24 NaHCO3, 0.5 Na2HPO4, 5 HEPES, 2.5 CaCl2, 1.2 MgCl2 + 0.1% BSA.
For glucagon secretion, batches of 10 islets were 1) preincubated in buffer supplemented with 10 mmol/L glucose, before incubation with 10 mmol/L glucose, 0.5 mmol/L glucose, or 0.5 mmol/L glucose + 5 μmol/L epinephrine (Sigma-Aldrich, cat. no. E4250) or 5 μmol/L GC for 1 h at 37°C, or 2) preincubated in 17 mmol/L glucose before incubation with 17 mmol/L glucose, 2 mmol/L glucose, or 2 mmol/L glucose + 5 μmol/L epinephrine. Glucagon released into the supernatant was then measured using homogeneous time resolved fluorescence (HTRF) ultrasensitive assay (Cisbio, cat. no. 62CGLPEG) or Lumit bioluminescent immunoassay (Promega, cat. no. CS3037A06) (26).
For insulin secretion, batches of 10 islets were preincubated in buffer supplemented with 3 mmol/L glucose before sequential incubation in 3 mmol/L glucose, 17 mmol/L glucose, and 17 mmol/L glucose + 10 mmol/L KCl or 5 μmol/L GC for 30 min at 37°C. Insulin was measured using HTRF ultrasensitive assay (Cisbio, cat. no. 62IN2PEG) or Lumit bioluminescent immunoassay (Promega, cat. no. CS3037A01). Total glucagon and insulin were extracted from islets lysed in acid ethanol.
α-Cells were identified in an unbiased manner by their characteristic Ca2+ spiking activity at low glucose, as well as responses to epinephrine, as reported (22,27). Signal contributions from β-cells are unlikely given that they are electrically silent at low glucose and inhibited by epinephrine. The proportion of low glucose–responsive α-cells was calculated as the area occupied by identified α-cells normalized to total islet area. Ca2+ spike amplitude was calculated for individual cells using Δ 340/385 nm or F/Fmin.
GC, F-actin, G-actin, and glucagon were analyzed using corrected total cell fluorescence (CTCF), as previously described. CTCF accounts for the effect of cell size on fluorophore intensity by taking the integrated density and subtracting area of the selected cell’s mean background fluorescence (28,29). α-Cell, β-cell, and δ-cell area and size were analyzed using the ImageJ (National Institutes of Health) Particle Analysis plugin applied to binarized and thresholded images. Linear adjustments to brightness and contrast were applied to representative images, with intensity values maintained between samples to allow cross-comparison.
Statistical details for each experiment are reported in the corresponding figure legend. The n number represents animal, batch of islets, or donor. Data normality was assessed using D’Agostino-Pearson test. All analyses were conducted using GraphPad Prism 9.0 software. Pairwise comparisons were made using the two-tailed unpaired t test (parametric) or Mann-Whitney test (nonparametric). To assess multiple interactions, one-way or two-way ANOVA were used, adjusted for repeated measures where needed. Post hoc comparisons were made using Bonferroni, accounting for degrees of freedom. Linear regression was used to assess strength of association between explanatory and dependent variables, with slopes compared using ANCOVA. Data represent mean ± SEM or SD, with individual data points shown where possible. Where a large number of data points obscure the mean ± SEM or SD, violin plots are instead used (showing median and interquartile range).
Mouse studies were regulated by the Animals (Scientific Procedures) Act 1986 of the U.K. (Personal Project Licenses P2ABC3A83 and PP1778740). Approval was granted by the University of Birmingham’s Animal Welfare and Ethical Review Body.
Human pancreas sections were obtained from Alberta Diabetes Institute IsletCore (30). Procurement of human pancreases was approved by the University of Alberta Human Research Ethics Board (Pro00013094). All donors’ families gave informed consent for the use of pancreatic tissue in research. Studies with human tissue were approved by the University of Birmingham Ethics Committee and by the National Research Ethics Committee (REC reference 16/NE/0107, Newcastle and North Tyneside, U.K.). Donor characteristics are reported in Supplementary Table 1. Anonymized donor IDs can be cross-referenced against the IsletCore database (www.isletcore.ca), including information about cold ischemia time, total islet equivalents isolated, tissue purity, insulin content, and stimulation index.
Data and Resource Availability
All data generated or analyzed during this study are included in the published article (and its online supplementary files). Source data files generated and/or analyzed during the current study are available from the corresponding author upon reasonable request. Noncommercially available reagents are available from the corresponding author upon reasonable request.
Deletion of GC Increases Basal Glucagon Secretion During the HFD
Mice with global GC deletion were used, since 1) GC/GC/Gc is exclusively expressed in α-cells and liver (18,22); 2) Gcg-Cre lines have variable recombination efficiency and specificity (31); 3) recently reported Gcg-CreERT2 knock-in mice require tamoxifen induction (32), which interferes with hepatic triglyceride accumulation and hence GC levels; and 4) two patients with homozygous inactivating mutations in GC have been described (33,34). The GC−/− mice used here are phenotypically well validated, do not possess detectable GC/Gc expression, and have 50% reduced and 90% reduced 25(OH)D and 1,25(OH)D levels, respectively (22,25).
GC−/− and littermate control GC+/+ mice were fed the HFD for up to 12 weeks, with glucose tolerance tested every 4 weeks. The GC+/+ cohort included some heterozygous (GC+/−) animals as controls, since we did not see any phenotypic differences versus wild-type mice (GC+/+). As expected from our previous studies, glucose tolerance was similar in GC+/+ and GC−/− mice fed the SC (i.e., 0 weeks HFD) (Fig. 1A and B). No significant differences were observed in glucose tolerance in GC+/+ and GC−/− mice at 4 weeks, 8 weeks, and 12 weeks of the HFD (Fig. 1C–H). Confirming efficacy of the preclinical obesity model, 4-week HFD-fed GC+/+ and GC−/− mice were glucose intolerant versus age-matched controls fed the SC (Fig. 1I). Body weight gain was similar in female and male GC+/+ and GC−/− mice during the HFD (Fig. 1J and K).
As a similar phenotype was observed in both female and male mice, we combined both sexes for subsequent studies. Plasma glucagon levels were assessed at 0 and 30 min post injection of insulin. While glucose levels were lowered to a similar extent in GC+/+ and GC−/− mice (Fig. 1L), basal fasted glucagon secretion was significantly (twofold) elevated in GC mice after 4 weeks of the HFD (Fig. 1M–O). Glucagon-to-glucose ratios, calculated using measures from the same animal, provided further evidence of dysregulated basal but not stimulated glucagon secretion (Fig. 1M and P).
In summary, GC−/− mice are glucose tolerant during the HFD, but display elevated basal glucagon levels, indicative of defective α-cell function.
HFD GC−/− Mice Have Aberrant α-, β- and δ-Cell Morphology
Pancreata isolated from HFD-fed GC+/+ mice showed a twofold increase in GC protein levels versus the SC controls (Fig. 2A and B). GC protein was undetectable in pancreata from HFD-fed GC−/− mice, further demonstrating the reliability of the antibody and immunostaining approach used (22) (Fig. 2A and B). We previously showed that pancreata from SC-fed GC−/− mice possess decreased α-cell mass and α-cell size (22). We thus performed high-resolution morphometric analysis in pancreata from HFD-fed mice.
The HFD feeding itself did not affect islet area occupied by α-cells or α-cell size compared with age-matched SC controls (Fig. 2C–E). However, a large reduction in α-cell size was observed in HFD-fed GC−/− mice versus GC+/+ littermates (Fig. 2C–E). By contrast to its effects on α-cells, the HFD increased β-cell size in GC+/+ islets, with a further increase detected in GC−/− islets (Fig. 2C, F, and G). Analysis of δ-cells revealed an HFD-induced increase in their proportion, a change that was partly reversed by deletion of GC (Fig. 2H–J).
In summary, these data suggest that, during HFD, GC restrains β-cell size, while promoting α-cell size and δ-cell expansion to support normal plasticity.
Glucagon Secretion and α-Cell Ca2+ Responses Are Impaired in HFD GC−/− Islets
Islets were isolated from HFD-fed GC+/+ and GC−/− mice and their age-matched SC controls for detailed in vitro analyses. As reported previously (22), SC GC−/− islets presented with impaired low glucose- and low glucose + epinephrine–stimulated glucagon secretion versus GC+/+ littermates (Fig. 3A). Similar impairments were detected for the HFD, although responses to epinephrine remained intact, suggesting that the defect is upstream of the exocytotic machinery (Fig. 3A). At a glucose concentration submaximal for α-cell function (i.e., 2 mmol/L), glucagon secretion was still reduced in HFD GC−/− islets (Fig. 3B). Glucose-stimulated insulin secretion tended to be increased in SC GC−/− islets, and this trend became significant during the HFD (Fig. 3C). A tendency toward increased basal glucagon secretion was also noted in HFD GC−/− islets (Fig. 3B), which might partly explain the increased insulin secretion, since glucagon is insulinotropic when β-cells are active (35). No significant differences in total glucagon or insulin content could be detected between GC−/− islets and GC+/+ controls (Fig. 3D and E).
Given the apparent changes in glucagon and insulin secretion, we next investigated upstream Ca2+ fluxes, with α-cells identified by their characteristic responses to low glucose (0.5 mmol/L) as well as epinephrine (27). Confirming our previous findings, proportion of active α-cells (i.e., percentage of cells displaying Ca2+ spikes; a measure of recruitment into Ca2+ activity) was decreased in SC GC−/− islets (Fig. 3F–I). By contrast to our previous results, we also observed a significant decrease in Ca2+ amplitude in SC GC−/− islets (Fig. 3G–I). The most likely explanation for this discrepancy is the relatively advanced age of the SC mice used in the study here, which were age-matched with those fed the HFD, and this suggests that age might exacerbate the in vitro phenotype following GC deletion. Nonetheless, the HFD decreased both the proportion of active α-cells as well as the amplitude of their Ca2+ spikes (Fig. 3F–I). The effect of the HFD on Ca2+ spike amplitude, but not proportion of active α-cells, was exacerbated following loss of GC (Fig. 3F–I) (Supplementary Movies 1 and 2). Ca2+ imaging results were validated using a second Ca2+ probe (Fluo 8), confocal microscopy, and a higher magnification objective (Fig. 3J and K) (Supplementary Movies 3 and 4).
GC-Dependent Actin Cytoskeleton Remodelling Occurs During the HFD
During stimulation, the F-actin cytoskeleton undergoes rearrangement to facilitate exocytosis of hormone vesicles (36–38). In line with the actin-scavenging function of GC, we previously showed that F-actin density was increased in GC−/− islets, leading to changes in glucagon granule morphology and distribution, suggestive of sequestration and trapping (22). Directly implicating the F-actin cytoskeleton in glucagon release, incubation of GC−/− islets with latrunculin B was able to restore function (22). We therefore investigated whether restoration of GC levels and, ergo, the F-actin cytoskeletal structure might rescue the phenotype of HFD GC−/− islets. Following acute (10 min) stimulation with low glucose, F-actin density was decreased in GC−/− islets but unchanged in islets from GC+/+ mice (Fig. 4A–C). F-actin remained low in GC−/− islets after chronic (60 min) stimulation but was increased approximately twofold in GC+/+ islets (Fig. 4A–C).
As we previously showed (22), deletion of GC from SC islets led to increased F-actin density concomitant with a decrease in G-actin monomers, presumably due to their involvement in forming polymerized actin (Fig. 4D–F). On the other hand, F-actin density and fiber thickness increased by almost threefold in HFD GC+/+ islets (Fig. 4D–F). Unexpectedly, given its actin-scavenging function, deletion of GC led to a decrease in F-actin density in HFD GC−/− islets versus GC+/+ controls (Fig. 4D and E). By contrast, monomeric G-actin was increased in HFD GC−/− islets, suggesting that G-actin is sequestered away from sites of F-actin polymerization following deletion of GC (Fig. 4D–F). In all cases, changes in F-actin and G-actin were detected throughout the islet (Fig. 4E and F) as well as in individual α-cells (Fig. 4G and H) and β-cells (Fig. 4I and J), suggesting that glucagon granule-resident GC acts in a paracrine manner to influence cytoskeletal structure throughout the islet (i.e., by severing and depolymerizing F-actin into G-actin) (22,38).
GC Supplementation Restores F-Actin Cytoskeletal Structure and Glucagon Release
We next investigated whether exogenous GC could modify F-actin levels in GC−/− islets to restore α-cell activity. Using a published RNA-seq data set (19), transcripts for the endocytic receptors responsible for GC uptake, megalin (Lrp2) and cubilin (Cubn) (14,39–41), were found to be expressed in purified α-cells at a similar level to the gastric inhibitory polypeptide receptor (Gipr) (normalized expression: 8.9 ± 5.3 vs. 9.7 ± 3.4 vs. 8.0 ± 7.0 for Lrp2 vs. Cubn vs. Gipr, respectively; taken from GSE76017). As expected from this, GC levels could be restored in HFD GC−/− islets following incubation with exogenous protein (Fig. 5A–C). Confirming the directionality of F-actin changes, treatment of HFD GC−/− islets with GC restored F-actin levels to wild-type levels (Fig. 5D), as seen throughout the islet as well as in individual α-cells (Fig. 5C–E).
As expected, low glucose (0.5 mmol/L)-stimulated glucagon secretion was impaired in HFD GC−/− islets (22). Pertinently, application of GC restored normal glucagon secretion in HFD GC−/− islets, without influencing the function of HFD GC+/+ islets (Fig. 5F). The effects of GC on glucagon secretion were not associated with increases in intracellular Ca2+ concentration, which was slightly but significantly decreased in GC-treated islets (Fig. 5G and H). Reflecting either the lowered glucagon tone or decreased F-actin in GC−/− islets, insulin secretion failed to shut off at low glucose (0.5 mmol/L) (Fig. 5I), an effect remarkably similar to that reported when the small guanosine-5′-triphosphatase and actin polymerizer, RhoA, is inhibited in α-cells (42). GC treatment was unable to restore this defect or influence basal insulin levels in GC+/+ or GC−/− islets (Fig. 5I). By contrast, GC treatment led to a large (∼10-fold) amplification of glucose-stimulated insulin secretion, with a greater effect in HFD GC−/− islets (Fig. 5J)
Finally, as a proof of principle, we were able to show that GC could be supplemented in human islets, leading to increases in glucagon granule area as well as F-actin density (Fig. 5K–M), visualized at ∼110-nm resolution using structured illumination microscopy.
GC Expression Is Decreased in Islets of T2D Donors
In pancreas sections from donors with no diabetes (ND), GC expression was only present in α-cells, as expected (18,22,43) (Fig. 6A). While a similar staining pattern was observed in pancreas sections from T2D donors, GC expression levels were approximately twofold reduced (Fig. 6A). Some interindividual variability was observed, but reduced GC expression appeared to be a remarkably consistent feature of T2D (Fig. 6B and C). Reflecting findings in HFD mice, analysis of individual α-cells in T2D donors revealed a decrease in cell size (Fig. 6D and E). While the proportion of islet area occupied by δ-cells was unchanged during T2D, δ-cell size was slightly but significantly reduced (Fig. 6F–I).
Linear regression showed a strong correlation between GC and glucagon expression in α-cells from ND donors (Fig. 6J). While a significant linear correlation was also detected for individuals with T2D, the strength of correlation was much lower (Fig. 6K), consistent with the reported decrease in GC expression (Fig. 6B) as well as in α-cell glucagon expression (Fig. 6L). As expected from this, the regression slopes were significantly different between ND and T2D samples (P < 0.001). Together, these analyses show that glucagon expression covaries with GC expression and that this relationship is partly lost during T2D.
In the current study, we show that deletion of GC in HFD-fed animals leads to basal hyperglucagonemia and impaired low glucose-stimulated glucagon secretion. These secretory defects are associated with changes in Ca2+ fluxes, α-cell, β-cell, and δ-cell size and mass, as well as F-actin and G-actin abundance. Function of α-cells can be restored in GC−/− islets by using exogenous GC, which is taken up into cells following culture. Lastly, islets from T2D donors show decreases in GC expression, with concomitant changes in α-cell and δ-cell size and mass. Together, these results expand our previous findings on GC by revealing its regulatory role in glucagon secretion during metabolic stress and further suggest that GC is a pivotal component of the α-cell phenotype in health and disease. While GC is a signature gene expressed in α-cells, the current study shows that α-cells also have the potential to acquire GC via megalin-mediated endocytosis. This raises the possibility that circulating levels of GC may also contribute to α-cell GC-actin dynamics and phenotype.
In vivo metabolic phenotyping demonstrated that GC+/+ and GC−/− mice possess similar glucose excursion curves in response to an intraperitoneal glucose injection. However, basal plasma glucagon concentrations were consistently raised in GC−/− mice, in line with a tendency toward elevated glucagon secretion from isolated islets at 17 mmol/L glucose, which would be expected to increase hepatic glucose output. One possible explanation for the apparently normal glucose homeostasis is that the increase in glucagon levels is not sufficient to influence insulin counterregulation, or might even act to prime β-cells for insulin secretion (44,45). Alternatively, recent studies have shown that Gc is upregulated in dedifferentiated β-cells and that deletion of Gc increases glucose-stimulated insulin secretion and liver insulin sensitivity at 12 weeks of the HFD (24). In any case, these data show that HFD-induced basal hyperglucagonemia (46) is further aggravated following GC deletion. We were unable to reliably detect significant increases in in vitro insulin secretion or β-cell function at 4–8 weeks of the HFD, arguing against this possibility here, although we concede that clamp studies are needed to properly assess this. Moreover, GC expression was variably upregulated in β-cells, remaining much lower than the levels seen in α-cells.
Plasma glucagon levels, stimulated by insulin injection, were almost identical in HFD-fed GC+/+ and GC−/− mice, despite impaired glucagon release from isolated islets incubated in low (0.5 mmol/L) glucose. We found, however, that the effect of GC deletion was milder in islets exposed to submaximal (2 mmol/L) glucose concentration, which would be closer to that achieved in vivo. Together, these data suggest that GC is relatively more important in α-cells operating close to their functional ceiling, with the caveat that in vitro glucagon secretion assays might be less sensitive at 2 mmol/L glucose due to the relatively smaller magnitude change. Along similar lines, HFD-fed GC−/− mice presented with basal hyperglucagonemia at blood glucose concentrations of 10–11 mmol/L, while in isolated islets basal glucagon secretion was similar in GC+/+ and GC−/− mice at high glucose. One explanation for this discrepancy might lie in the finding that glucose-stimulated insulin secretion was increased twofold in islets of HFD-fed GC−/− mice. In vivo, relative hyperinsulinemia would be expected to drive hyperglucagonemia to maintain blood glucose levels, which were not different between HFD-fed GC+/+ and GC−/− mice. Another explanation might lie in the changes in α-cell morphology observed in pancreas sections taken from HFD-fed GC−/− animals. A decrease in α-cell size might lead to an increase in α-cell membrane juxtaposed with the islet capillaries, favoring release of glucagon into the circulation.
During the SC diet, we showed that loss of GC leads to a large increase in the density of the F-actin cytoskeleton (and concomitant decrease in G-actin), acting as a physical barrier against exocytosis of glucagon granules during low-glucose stimulation (22). Changes in the F-actin and G-actin cytoskeleton occur throughout the islet, since ∼50% of GC is present in glucagon granules and can readily be taken up by neighboring cells by endocytosis, as shown here following application of exogenous GC. Following 4–8 weeks of the HFD, F-actin density was increased almost twofold in GC+/+ islets. On a background of metabolic stress, deletion of GC did not further increase F-actin density. In fact, HFD GC−/− islets showed a surprising reduction in F-actin density, contrary to our previous findings in standard diet islets (22). A similar decrease in F-actin density was seen in GC−/− islets stimulated with low glucose for 60 min. Notably, treatment with exogenous GC replenished intracellular GC and F-actin levels in HFD GC−/− islets, confirming that GC acts to increase F-actin density during metabolic stress. Given that GC is a potent actin scavenger, what might be the mechanisms involved in this apparent decrease in F-actin? A likely mechanism revolves around G-actin, which was virtually undetectable in HFD GC−/− islets. Without G-actin to supply monomers, polymerized F-actin cannot be formed. Indeed, previous reports by us have shown a similar decrease in F-actin in trophoblasts depleted for GC, which was associated with an increase in G-actin monomers in the nucleus where they are unavailable for assembly into polymerized F-actin (39). Another mechanism might be a large compensatory upregulation in gelsolin, which severs F-actin into G-actin (36,47), although we would expect this to be associated with an increase in G-actin levels.
Kuo et al. (24) recently reported that GC−/− islets show an insulin signaling/sensitivity defect but exhibit a normal glucagon phenotype under both a standard diet and HFD conditions. Since our studies used islets from animals fed the HFD for 4 and 8 weeks, we cannot exclude that glucagon secretion in GC−/− animals/islets normalizes in line with improved β-cell function at 12 weeks, the feeding period used by Kuo et al. (24). We also used a different GC-knockout mouse line, which might give rise to a different phenotype. However, it should be noted that these animals are well validated by multiple groups and show complete loss of GC in α-cells and the liver, undetectable circulating GC based on liquid chromatography–mass spectrometry, and a 90% reduction in circulating 25(OH)D in homozygotes (22,25). Suggesting that GC plays a critical role in α-cell biology: 1) GC/Gc is an α-cell signature gene; 2) GC protein expression is upregulated during the HFD, remaining 10- to 100-fold higher than that in β-cells; and 3) defects in α-cell function have a clear mechanistic basis, including changes in cell morphology, cell mass, cytoskeletal structure, and ionic fluxes, shown also by recent Patch-seq studies (8). Nonetheless, these studies together posit that, depending on the duration of metabolic stress, effects of GC deletion (and GC supplementation) can be seen on both the insulin and glucagon axes. Further studies using conditional GC deletion in α-cells and β-cells are warranted.
Studies in human donor pancreas sections revealed that GC and glucagon expression are positively associated, with levels covarying across hundreds of individual cells examined, a relationship that was lost during T2D. Mechanistically, this observation likely reflects changes in the α-cell transcription factor network, since GC possesses cell type-selective open chromatin regions (18). Ultimately, however, altered gene regulation must impact functional protein targets, and our in vitro findings support the notion that the disrupted relationship between GC and glucagon might contribute to impaired glucagon secretion during T2D. Further studies are warranted in isolated human islets to investigate the effects of silencing GC on glucagon expression and secretion in α-cells.
Changes in the actin cytoskeleton could be rescued using exogenous GC. In the kidney, GC-bound 25(OH)D is taken up by facilitated endocytotic uptake via the megalin-cubulin complex (14,41,48), where liberated 25(OH)D is then converted to 1,25(OH)2D. Immunostaining clearly showed dose-dependent uptake of GC into islets, demonstrating that similar transport mechanisms also exist in the pancreas, as suggested by published RNA-seq data (49). These data suggest that, unusually, decreases in expression of a key cell signature gene can be offset by supplementing its protein product and warrant further investigation of the uptake mechanisms involved. While these results point to GC as a therapeutic target, caution should be extended due to opposing effects of GC on both the α-cell and β-cell compartments (24). However, it should be noted that high glucose levels have been shown to inhibit megalin-mediated endocytosis, which might differentially affect GC uptake into α-cells and β-cells (50). Moreover, molecular addresses such as V1BR could be used to target GC specifically to α-cells (51–53). In any case, we envisage that GC administration during T2D could maintain α-cell function, while restraining β-cell proliferation and hyperinsulinemia, known to drive insulin resistance (54,55). Our data in human pancreas sections support a reduction in GC during T2D, lending further weight to this argument. Nonetheless, careful preclinical studies in mice at various time points are required to assess this.
The current study has a number of limitations. First, we used a well-phenotyped global GC−/− mouse model in which GC is undetectable in the circulation. While GC−/− mice are vitamin D sufficient (22), we cannot exclude that loss of circulating GC influences α-cell phenotype and function. In the future, it will be worthwhile conditionally deleting GC in the α-cell or liver to explore the role of circulating GC in the α-cell and, more widely, islet function.
Second, we decided to investigate HFD at 4 and 8 weeks, since longer feeding periods did not lead to further changes in glucose tolerance. In any case, this length of the HFD allowed α-cell function to be determined without any confounding affects caused by GC upregulation in the β-cell compartment, as shown by our immunohistochemical analyses.
Third, the mild in vivo phenotype seen in HFD-fed GC−/− mice might reflect compensation, especially since the gene was (presumably) deleted during development.
Finally, while GC supplementation increased glucagon granule area/density in human α-cells, it should be noted that the decrease in GC levels seen in samples from T2D donors is associative and might be a consequence rather than a cause of changes in α-cell morphology and function.
In summary, we show that α-cells lacking GC fail to adapt properly to metabolic stress, displaying a range of defects leading to impaired basal and low glucose-stimulated glucagon release. Given its role under both normal and obesogenic conditions, GC should thus be considered as a key regulator of α-cell function and glucagon secretion.
This article contains supplementary material online at https://doi.org/10.2337/figshare.21553317.
Acknowledgments. The authors thank Prof. Patrick E. MacDonald and Dr. Jocelyn Manning Fox (Alberta Diabetes Institute IsletCore at the University of Alberta in Edmonton) for provision of human pancreas sections.
Funding. D.J.H. was supported by the Medical Research Council (MR/S025618/1) and Diabetes UK (17/0005681) Project Grants, as well as a UK Research and Innovation European Research Council Frontier Research Guarantee Grant (EP/X026833/1). This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 Research and Innovation Programme (Starting Grant 715884 to D.J.H.) and the Department of Health. L.J.B.B. was supported by the Wellcome Trust Sir Henry Wellcome Postdoctoral Fellowship (201325/Z/16/Z) and a Trinity College, Oxford Junior Research Fellowship. The research was funded by the National Institute for Health Research (NIHR) and Oxford Biomedical Research Centre (BRC). Human pancreas sections were provided by the University of Alberta in Edmonton Alberta Diabetes Institute IsletCore (https://www.bcell.org/adi-isletcore.html) with the assistance of the Human Organ Procurement and Exchange (HOPE) program, Trillium Gift of Life Network (TGLN), and other Canadian organ procurement organizations.
The views expressed are those of the author(s) and not necessarily those of the National Health Service, the National Institute for Health Research, or the Department of Health. The funders had no role in study design, data collection, data analysis, interpretation, or writing of the paper.
Duality of Interest. D.J.H. receives licensing revenue from Celtarys Research. K.V., M.H., and D.J.H. are named on a patent application regarding the use of GC-globulin as an antidiabetes therapy. No other potential conflicts of interest relevant to this article were reported.
Author Contributions. K.V., D.N., J.A., A.H., F.C., S.H., and L.J.B.B. performed experiments and analyzed data. K.V., M.H., and D.J.H. wrote the paper with input from all authors. M.H. and D.J.H. conceived and designed the studies. M.H. and D.J.H. supervised the studies. M.H. and D.J.H. are the guarantors of this work, and, as such, had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.